Comparative Health Effects in Mice of Libby Amphibole Asbestos and a Fibrous Amphibole from Arizona

Toxicology and Applied Pharmacology 334 (2017) 24–34

Contents lists available at ScienceDirect

Toxicology and Applied Pharmacology

journal homepage: www.elsevier.com/locate/taap

Comparative health eﬀects in mice of Libby amphibole asbestos and a MARK ﬁbrous amphibole from Arizona

⁎ Jean C. Pfaua, , Brenda Buckb, Rodney V. Metcalfb, Zoie Kaupisha, Caleb Staira, Maria Rodrigueza, Deborah E. Keila a Department of Microbiology and Immunology, Montana State University, Bozeman, MT 59717, USA b Department of Geoscience, University of Nevada Las Vegas, Las Vegas, NV 89154, USA

ARTICLE INFO ABSTRACT

Keywords: This project developed from studies demonstrating that Libby Amphibole Asbestos (LAA) causes a non-typical set Amphibole of health outcomes not generally reported for asbestos, including systemic autoimmunity and an unusual and Asbestos devastating lamellar pleural thickening that progresses to severe pulmonary dysfunction and death. Further, Autoantibodies mineral fiber mixtures with some similarities to LAA have recently been discovered in southern Nevada and Lung fibrosis northwestern Arizona, where the material exists in extensive recreational areas and is present in yards, roads, Pleural fibrosis parking lots and school yards. The objective was to compare the health outcomes in mice exposed to either LAA or the fibrous amphiboles collected in Arizona at the Lake Mead National Recreational Area at very low doses to represent environmental exposures. In this study, the fibrous amphibole asbestos sample from Arizona (AzA) is composed of winchite (69%), actinolite (22%), and non-amphibole minerals (9%) and has a mean aspect ratio of 16.7 ± 0.9. Fibrous amphibole asbestos from Libby (LAA) is composed of winchite (70%), richterite (9%), tremolite (5%), and non-amphibole minerals (16%) with a mean aspect ratio of 8.4 ± 0.7. C57BL/6 mice were exposed by oropharyngeal aspiration to fiber suspensions at a very low dose of 3 μg/mouse. After seven months,

both LAA- and AzA-exposed mice had indices of chronic immune dysfunction related to a TH17 cytokine profile, with B cell activation, autoantibody production and proteinuria, suggesting kidney involvement. In addition, both exposures led to significant lung and pleural fibrosis. These data suggest that there is risk of pulmonary disease and autoimmune outcomes with environmental exposure to amphibole asbestos, and that this is not limited to Libby, Montana.

1. Introduction mice, providing a critical tool for evaluation of the relative toxicity of other mineral fibers (Blake et al., 2008; Ferro et al., 2013; Gilmer et al., Multiple studies have chronicled the devastating health outcomes 2015; Pfau et al., 2013; Pfau et al., 2008; Zebedeo et al., 2014). that resulted from asbestos exposure in Libby, Montana. While the rates LAA exposures occurred due to contamination of vermiculite, which of pulmonary fibrosis (asbestosis) and cancer (mesothelioma, pul- was mined outside of Libby for decades and used throughout the monary carcinoma) are significantly elevated among people exposed to community in buildings, gardens, and playgrounds. This meant that the Libby amphibole asbestos (LAA), the predominant negative health there was a wide range of exposures, from high occupational exposures outcomes include systemic autoimmunity and a progressive pleural fi- to relatively low, environmental exposures (Noonan, 2006; Noonan brosis that may also be autoimmune in nature through production of et al., 2015). Recently, the Environmental Protection Agency (EPA) mesothelial cell autoantibodies (MCAA) (Peipins et al., 2003; Pfau conducted a risk assessment specifically for LAA based on a study that et al., 2005; Rohs et al., 2008; Sullivan, 2007; Szeinuk et al., 2016; U.S. showed that significant negative health effects were occurring at ex- Environmental Protection Agency, R, 2011; Whitehouse et al., 2008; tremely low exposure levels (Lockey et al., 2015). LAA is the first as- Winters et al., 2012; Larson et al., 2010a; Gilmer et al., 2016; Hanson bestiform fiber for which a toxicity value (Reference concentration, et al., 2016; Marchand et al., 2012; Serve et al., 2013). Using a wildtype RfC) has been derived to help define a remediation target that reduces mouse model (C57BL/6), we have corroborated these outcomes in to acceptable levels the risk of acquiring non-malignant respiratory

Abbreviations: ANA, antinuclear autoantibodies; AzA, Arizona amphibole asbestos; EDS, energy dispersive spectroscopy; EPMA, electron probe microanalysis; LAA, Libby amphibole asbestos; LERP, Libby Epidemiology Research Program; MCAA, Mesothelial Cell Autoantibodies; RfC, Reference Concentration; SAED, selected area electron diﬀraction ⁎ Corresponding author at: Department of Microbiology and Immunology, 960 Technology Blvd, Rm120, Bozeman, MT 59718, USA. E-mail addresses: [email protected] (J.C. Pfau), [email protected] (B. Buck), [email protected] (R.V. Metcalf), [email protected] (D.E. Keil). http://dx.doi.org/10.1016/j.taap.2017.08.022 Received 14 June 2017; Received in revised form 26 July 2017; Accepted 31 August 2017 0041-008X/ © 2017 Elsevier Inc. All rights reserved. J.C. Pfau et al. Toxicology and Applied Pharmacology 334 (2017) 24–34

disease. The RfCLAA was released on December 8, 2014 at 0.00009 PCM Table 1 f/cm3 (Phase Contrast Microscopy fibers/cm3)(U.S. Environmental Fiber characteristics. Protection Agency, R, 2015). The dramatic outcome of this fiber-spe- fi fi Characteristics LAA (whole AzA (amphibole AzA (whole ci c health assessment, based on a speci c non-cancer outcome (pleural sample)a,b only) sample)b fibrosis), emphasizes the need to evaluate health impacts based on variable mineral fiber composition from different sites. Similar to the Mineralogy Winchite (70%), Winchite (76%), Winchite (69%), study used by the EPA, the Libby Epidemiology Research Program Richterite (9%), Actinolite (24%) Actinolite (22%), Tremolite (5%) Non-Amphibole (LERP, ATSDR, TS000099-01) has shown that 50% or more of people Non-Amphibole (9%) ff exposed to Libby Amphibole su er from pleural scarring, and that this (16%) scarring can dramatically impact pulmonary function, eventually leading to significant disability and death (Szeinuk et al., 2016; Black Morphology Analytical TEM SEM SEM et al., 2014). Amphibole mineral fibers are found in soils and rock method outcroppings in many parts of the U.S. and around the world, leading to Min width (μm) 0.05 0.2 0.2 human exposures to naturally occurring asbestos (NOA) through nu- Max width (μm) 3.0 17.5 45.6 merous routes, including land development, recreation, and use of the Mean width (μm) 0.36 ± 0.01 0.7 ± 0.1 1.3 ± 0.2 μ material in roads, parking lots, and playgrounds (Abakay et al., 2016; Min length ( m) 0.2 1.0 1.0 Max length (μm) 43.6 151.0 151.0 Bayram & Bakan, 2014; Carbone et al., 2016; Cooper et al., 1979; Mean length 2.3 ± 0.2 9.0 ± 0.7 9.7 ± 0.7 Environmental Protection Agency, US, 2008; Paoletti et al., 2000; Van (μm) Gosen, 2007; Wylie & Candela, 2015). Thus, the public health impacts Min aspect ratio 1.0 1.0 1.0 of on-going, current exposures, in addition to exposures over the last Max aspect ratio 145.3 263.4 263.4 Mean aspect ratio 8.4 ± 0.7 18.2 ± 1.1 16.7 ± 0.9 few decades, could be tremendous. Land use decisions in areas where Number of 510 427 470 fi NOA bers are discovered need to be based on strong data that can be particles used in evaluating the application of the very low Libby RfC more broadly to other amphibole NOA. a Data from (Duncan et al., 2014; Lowers et al., 2012; Meeker et al., 2003). b Amphibole asbestos (fibrous amphiboles) has been reported in Used in the current study. rocks, soils, dust, and air from areas in southern Nevada and north- fi western Arizona on either side of the Colorado River near Hoover Dam used dental tools to separate AzA bers in the vein from the rock ma- (Buck et al., 2013; Metcalf & Buck, 2015; Tetra Tech, I, 2014). The trix. This method decreases the amount of contaminant accessory mi- fi amphibole asbestos material exists in extensive recreational areas and is nerals such as quartz, feldspar, and mica, which co-occur with the - present in yards, roads, parking lots and school yards in and around brous amphiboles. Scanning electron microscope-energy dispersive Boulder City, Nevada (Buck et al., 2013; Metcalf & Buck, 2015). These spectroscopy (SEM-EDS) analyses were performed on 470 particles to materials continue to be disturbed both by road construction/urban measure particle size and shape, and mineral chemistry. Wavelength development but also by natural desert processes that produce dust dispersive electron probe microanalysis of a polished thin section of the storms throughout the year. Human exposures may already be ex- AzA vein material provided quantitative chemical analyses that were tensive. The pathogenicity of this material is unknown, although en- used to classify the AzA minerals, methods comparable to that used by vironmental exposure to asbestos in southern Nevada is supported by a the USGS to classify LAA samples (Meeker et al., 2003). Characteristics fi fi study showing atypical distribution of mesothelioma among women of these bers are summarized in Table 1. The bers used in the current and young people (Baumann et al., 2013; Baumann et al., 2015). Am- study are shown in columns 1 and 3 of Table 1. fi ff phibole asbestos in southern Nevada is predominantly the regulated All bers were suspended in sterile phosphate bu ered saline (PBS, mineral actinolite, while on the Arizona side the dominant asbestos pH 7.4), and sonicated (Branson Ultrasonics, Danbury, CT) for 5 min fi mineral, referred to here as Arizona amphibole asbestos (AzA), is prior to use to minimize aggregation of the bers. fi winchite, and is similar in composition and morphology to LAA. The Endotoxin testing of the ber suspensions was performed using the ® relative pathogenicity may determine the need for a public health risk PyroGene Recombinant Factor C Endotoxin Detection System assessment for both cancer and non-cancer outcomes in this region. (Cambrex Bioscience, Walkersville, MD), following the manufacturer's This study was designed to initiate this critical area of study, using our protocol. Tested samples included the sterile PBS, suspended LAA (1.0 well-characterized mouse model. and 0.1 mg/ml) in sterile PBS, suspended AzA (1.0 and 0.1 mg·ml) in sterile PBS, plus E. coli O111·B4 (Sigma, St Louis, MO) lipopoly- 2. Materials and methods saccharide (LPS)-spiked samples, all against a standard curve provided with the kit. Briefly, all prepared samples and standards were placed in μ 2.1. Amphibole minerals a 96-well plate at 100 l/well, and incubated for 10 min at 37 °C. Detection working reagent was prepared and then added to all wells, fl LAA was provided by the EPA as a composite sample of asbestos- and the plate was read immediately on a uorescence plate reader rich rock samples collected from multiple sites in the W.R. Grace mine (Fluorskan Ascent FL, ThermoFisher) at 380 excitation/440 emission. outside of Libby, Montana. The LAA sample was previously character- The plate was incubated for an hour at 37 °C, then read again with the ized using a suite of methods including transmission electron micro- same settings using the Ascent Software, to allow calculation of the fl Δ scopy (TEM) with selected area electron diffraction (SAED) (Duncan change in uorescence ( RFU), which was then plotted against the et al., 2014); scanning electron microscopy (SEM) with energy dis- standard curve. Endotoxin was not detected in any of the samples, with persive spectroscopy (EDS) (Gavett et al., 2016), x-ray diffraction a detection limit of 0.01 EU/ml. (Lowers et al., 2012), and wavelength dispersive electron probe microanalysis (EPMA) (Meeker et al., 2003). 2.2. Mice and exposures The fibrous amphiboles in this Arizona-Nevada region originated by hydrothermal alteration of granitic host rocks where asbestos minerals All experiments with mice were approved by the Montana State were precipitated as fracture-fill veins. In order to produce an AzA University Institutional Animal Care and Use Committee (IACUC). The sample of high purity, a single sample of AzA in a centimeter-wide, mice used were wild type C57BL/6 (Charles River, Seattle, WA) asbestos-rich vein was collected from granitic rock of the Wilson Ridge maintained in the Montana State University Animal Resource Center. pluton in the Lake Mead Recreation Area in northwestern Arizona. We These mice were housed under specific pathogen free (SPF) conditions

25 J.C. Pfau et al. Toxicology and Applied Pharmacology 334 (2017) 24–34 with 12 h light-dark cycle, constant temperature (22 °C) and humidity 2.5. Flow cytometry (45%), and ad libitum food and water. Both male and female mice were used between 6 and 8 weeks of age. Baseline cheek bleeds were used to Cells were stained for 30 min at 4 degrees C, and then washed with evaluate autoantibody status at the beginning of the study, and initially, 1 ml ice cold PBS, twice. Staining was analyzed on a FACS Calibur flow no mice were positive for antinuclear autoantibodies (ANA, methods cytometer using Cell Quest software (BD Biosciences). Isotype control described below). antibodies (BD Biosciences) determined background staining, and < 1% of these controls was allowed in the M1 gate for percent positive. Monocytes and Lymphocytes were determined based on forward and 2.3. Oropharyngeal instillations side scatter, and then the major subsets within the lymphocyte gate were identified based on the antibodies listed above. Tight polygonal Sterile suspensions of the fibers were prepared at 0.1 mg/ml in regions were used (instead of quadrants) to identify the ultimate sub- sterile saline. Mice were briefly anesthetized with isoflurane until populations. breathing was slow and regular, but the animals were non-responsive. The tongue was pulled gently out of the mouth and to the side, and 2.6. Th1/Th2/Th17 cytometric bead array (CBA) 15 μl were injected into the oropharynx using a micropipette (1.5 μg/ mouse). Successful instillation was indicated by a deep breath of in- The cytometric bead array (CBA) mouse Th1/Th2/Th17 cytokine kit halation, and recovery from anesthesia was monitored. This procedure (BD Biosciences) examines seven cytokines in a single sample using was repeated 2 weeks later, for a total dose of 3 μg/mouse. This dose fluorescent beads that can be analyzed using flow cytometry. These was chosen to be within the range of environmental concentrations cytokines include interleukin-2 (IL-2), IL-4, IL-6, IL-17, IL-10, inter- recorded in Libby, MT, since the main concern for exposure to “natu- feron-γ (IFN-γ), and tumor necrosis factor-α (TNF-α), and are asso- rally occurring asbestos” (NOA) in Nevada and Arizona would be en- ciated with one of the three T-helper cell profiles, Th1, Th2, and Th17. vironmental (not occupational). Environmental exposures tend to be The cytokines belonging to the Th17 profile have been associated with sporadic, due to dust storms or recreational activity in areas with NOA, autoimmune diseases (Afzali et al., 2007; Furuzawa-Carballeda et al., rather than regular, all day, work exposures. In previous mouse ex- 2007). The CBA procedure was carried out according to the manufac- posure studies, a dose of LAA was chosen to represent cumulative oc- turer's instructions. To prepare the top standard, all seven of the lyo- cupational exposures in humans; but for this new study, a dose for LAA philized standard spheres were added to a 15 ml conical tube. The and AzA was selected to represent an environmental exposure. spheres were dissolved in 2.0 ml of the kit's assay diluent and incubated A second cheek bleed was performed 4 months after the second at room temperature for at least 15 min. The top standard was serially instillation. Animal weights were monitored weekly. Initially, 20 mice diluted to create the standard curve. A negative standard was also used were to be instilled in each group, but mice were lost due to death which only contained assay diluent. during instillation or during the 7 months due to excessive barbering Creating the mixture of capture beads was done by calculating the that required euthanasia, which is not unusual with this strain. Ending appropriate amount of each bead solution to use based on the number mouse numbers were 14 saline control, 9 LAA and 17 AzA exposed mice of samples for that particular experiment. All seven of the bead solu- that survived to the end of the experiment. 1 week prior to the end of tions were added to a single 5 ml centrifuge tube. Fifty microliters of the experiment, urine was collected from each mouse and assessed for the capture bead mixture were added to each 1.5 ml micro centrifuge proteinuria using albumin test strips (Albustix, Bayer Corp., Elkhart, assay tube. Another 50 μl of the standards and samples were added the IN). At 7 months after the last instillation, the mice were then eu- appropriate assay tube. Fifty microliters of PE detection reagent from thanized by CO2 asphyxiation, using IACUC-approved methods. the kit were added to all the assay tubes and incubated in the dark for 2 h at room temperature. One milliliter of wash buffer from the kit was added to all the assay tubes and then centrifuged at 200 g for 5 min. 2.4. Tissues harvested The supernatant was carefully removed still leaving some fluid to en- sure the pellet was not discarded. The pellet was resuspended in 300 μl After euthanization, blood was collected by cardiac puncture, al- of wash buffer. All samples were analyzed on the FACS Calibur flow lowed to clot at room temperature, and then centrifuged to collect cytometer for acquisition and analysis following instrument set-up serum. Serum was stored at −20 °C until use. 5 ml of sterile PBS with using the provided instrument set-up bead sets. 5% fetal bovine serum (FBS) were injected into the intact peritoneal cavity, gently agitated, and then removed using an 18-gauge needle. 2.7. Anti-nuclear antibody (ANA) assay This peritoneal wash fluid was centrifuged, and the cells were prepared for flow cytometry. Also harvested were lungs, diaphragms, and The serum samples from cardiac punctures were analyzed for anti- spleens. Lungs were weighed and fixed in Histochoice (Amresco, Solon, nuclear antibodies (ANA). The ANA test (ImmunoConcepts, OH), and the diaphragms were fixed in Histochoice lying flat in 12-well Sacramento, CA) is a semi-quantitative analysis that uses indirect im- tissue culture plates. The spleens were weighed, and then minced and munofluorescence to detect any auto-reactive antibodies in the serum. prepared as single cell suspensions as previously described (Ferro et al., The assay kit was modified for the use of mouse serum. 2013), including a brief wash in Red Blood Cell Lysis solution The serum was diluted at a 1:80 ratio by adding 5.0 μl of serum in (eBioscience/ThermoFisher; Waltham, MA) to select for white blood 395 μl of PBS. Twenty microliters of each diluted sample were added to cells. The splenocytes were counted using a Z-series Coulter Counter their corresponding wells on the HEp-2000 substrate slide and in- (Beckman Coulter, Brea, CA), and 1 million cells from each sample were cubated at room temperature for 30 min. The slides were washed with placed in two separate tubes containing 100 μl PBS with 3% Bovine PBS being careful not to cross contaminate any samples. The slide was Serum Albumin as blocking agent. Sets of cells were stained with BD then submerged in a staining dish with PBS for 10 min with gentle Biosciences (San Jose, CA) antibodies as follows: rotation. The slide was carefully dried by blotting in between the wells. B cells: CD19 (PE) or IgM (PerCP Cy5.5) The fluorescent antibody reagent was prepared by adding 2.0 μlof B1a B cells: IgMpos (PerCP Cy5.5), CD5pos (APC), CD23neg (PE) goat anti-mouse IgG antibody conjugated to AlexaFluor 488 B suppressor cells: CD19pos (PE), CD5pos (APC), CD1dpos (FITC), (Invitrogen, Eugene, OR) to 1.0 ml of PBS. Twenty microliters of the CD11bpos (PerCP Cy5.5) antibody solution was added to each well and incubated in the dark at T regulatory cells: CD3pos (PerCP Cy 5.5), CD4pos (PE), CD25pos room temperature for 30 min. The slide was washed as above and a (FITC), Fox P3pos (APC) cover slip was applied by placing four to five drops of Fluorsave

26 J.C. Pfau et al. Toxicology and Applied Pharmacology 334 (2017) 24–34

(Calbiochem, La Jolla, CA) along the midline of the slide then setting determined by quantitative EPMA analyses (Meeker et al., 2003). Of the cover slip. The slides were viewed using the FITC (488 nm) filter on just the AzA amphibole particles alone, 93% had aspect ratios 3.0 or a Nikon Eclipse 80i fluorescence microscope, using the 20× and 40× greater, and the mean aspect ratio was 18.2 ( ± 1.1) (Table 1, column objectives. Positive and negative controls were included on each slide, 2). Of the whole AzA sample with contaminant minerals included, 86% and slides were read by two trained, blinded, independent readers to of the particles have aspect ratios 3.0 or greater, and the mean aspect determine whether the sample was positive for ANA and to record ratio was 16.7 ( ± 0.9) (Table 1, column 3). The LAA sample overall staining patterns. has lower aspect ratios, with only 49% of particles having an aspect ratio 3.0 or greater. The mean width of the AzA particles was 1.3 2.8. Detection of mouse serum mesothelial cell autoantibodies (MCAA) ( ± 0.2) μm, which is essentially 3.6 times that of the LA, which was 0.36 ( ± 0.01) μm. This difference is artificial and reflects the different A cell-based ELISA was performed as previously described (Gilmer instrumentation used to characterize the fibers: the TEM is able to et al., 2015) to test for the binding of serum antibodies to mesothelial measure and identify significantly smaller particles than the SEM cells. Briefly, mouse primary mesothelial cells were seeded at con- (Table 1). Although the AzA sample contains numerous tiny fibers, they fluency on 96 well plates, attached overnight and fixed in 1% paraf- were too small for the current instrumentation (SEM/EDS) to reliably ormaldehyde. Following washing with PBS-Tween (0.05%), cells were measure and therefore those small particles were not included in the blocked with 5% non-fat dry milk/PBS and then exposed to serum from analyses. On-going analyses using field-emission scanning electron each mouse diluted in 3% BSA/PBS (1:100). Following a 2-h incubation microscopy is planned in order to more fully characterize this sample. with serum, cells were washed and blocked a second time. The secondary antibody HRP-conjugated goat anti-mouse IgG (Invitrogen) was 3.2. Mouse health and weights applied at a dilution of 1:1000 in 3% BSA/PBS and incubated for 1 h. Excess antibody was removed and plates developed using TMB reagent There was no significant difference in the average weights of com- (Thermo Scientific) followed by 50 μl 1 M HCl. Plates were analyzed at bined male and female mice in each treatment group (data not shown, 450 nm on a SpectrMax microtiter plate reader (Molecular Devices, p = 0.2). As expected for this strain, male mice tended to be larger than Sunnyvale, CA). Correction for non-specific secondary antibody binding females. The LAA-exposed females were slightly larger than saline fe- was performed on a plate-to-plate basis by subtracting the mean optical males, but overall, weights did not change with treatment (Fig. 1). density (OD) for the secondary antibody-only control wells from the There were no obvious changes in health status over the 7 months in mean OD of each sample. Samples were determined to be MCAA-po- any group. Mice in all 3 groups lost fur particularly on their faces and sitive (MCAA+) if the corrected OD was at least three standard de- necks, but without skin lesions. No tumors were found during tissue viations above the mean OD for control (normal, untreated) mouse harvest. serum. 3.3. Spleen weights and cell counts 2.9. QuickZyme hydroxyproline assay (right lung and diaphragm collagen) There was no significant difference in the average spleen weights for The amount of collagen in the right lung and diaphragm of mice was the different treatments (Fig. 2, top). Although the splenocyte counts determined by QuickZyme Total Collagen Assay (QuickZyme for LAA and AzA-treated mice were slightly higher than controls, these Biosciences, Leiden, The Netherlands) according to the manufacturer's were not significant at an alpha level of 0.05 (Fig. 2, bottom). When instructions. Right lungs were collected and weighed. Two same-sized corrected for body weight, AzA-treated female (but not male) mice (4 mm) tissue punches were obtained from the diaphragm of each showed increased numbers of splenocytes/g of mouse (Table 4). mouse. The right lungs and the combined tissue punches from each mouse were hydrolyzed overnight in 0.5 ml or 100 μl (respectively) of 3.4. Antinuclear autoantibodies (ANA) and mesothelial cell autoantibodies 6 M HCl at 95 °C and then the hydroxyproline was oxidized. After (MCAA) staining the hydroxyproline residues with QuickZyme stain, absorbance was measured at 540 nm on a Spectra-Max microtiter plate reader. At 4 months, LAA did not induce ANA above controls, but the 2.10. Statistical analyses

All graphs are representative of at least two experiments. Overall statistical differences were assessed by one-way ANOVA using StatPlus with Bonferroni post hoc testing (StatPlus Software, Walnut CA). Two- tailed t-tests were used to assess differences between two means, using StatPlus. Statistical significance was defined as p values < 0.05. Except for percentages (frequency), data are graphed with error bars indicating standard error of the mean.

3. Results

3.1. Fiber mineralogy and morphology

Morphology and mineralogy of LAA and AzA are shown in Table 1. Both whole samples have similar mineralogy (Table 1, columns 1 and 3). The LAA sample is reported to be 84% amphiboles, with the re- Fig. 1. Mouse weights at 7 months post-exposure. As expected, male mice tended to be maining 16% composed of various minerals including feldspar, quartz, larger than females. The LAA-exposed females were slightly larger than saline females. talc and calcite (Lowers et al., 2012). The AzA sample is composed of Saline: n = 14 (9 male, 5 female), LAA: n = 9 (4 male, 5 female), AzA: n = 17 (8 male, 9 91% amphiboles and 9% other accessory minerals including quartz and female). feldspar. Amphibole mineralogy is similar for both AzA and LAA sam- a = p < 0.05 compared with males in same treatment group; b = p < 0.05 compared ples: both are dominated by winchite (AzA 69%, LAA 70%) as to saline, same sex. Error bars = standard error of the mean (SEM).

27 J.C. Pfau et al. Toxicology and Applied Pharmacology 334 (2017) 24–34

Fig. 2. Spleen weights (top) and cell counts (bottom). There were no significant differences in spleen weights or cell counts among the treatment groups. Saline: n = 14 (9 male, 5 female), LAA: n = 9 (4 male, 5 female), AzA: n = 17 (8 male, 9 female). Error bars = standard error of the mean (SEM). percent positive for AzA was increased above saline controls (Fig. 3, top). By 7 months, both LAA and AzA had induced ANA, although LAA did not reach statistical significance at p < 0.05 (p = 0.1 by Fisher's Exact Test). Most of the positive ANA tests were from the female mice (Fig. 3, bottom). MCAA were induced primarily by AzA at this dose, shown both by percent positive (Fig. 4, top) and MCAA score (a measure of binding absorbance) (Fig. 4, bottom). Again, most of the positive tests were in female mice (Table 5). Despite what appears to be increasing trends with dose in the percent positive and scores in both sexes, neither sex fi Fig. 3. ANA testing was done on commercial slides from ImmunoConcepts at 1:80 serum met statistical signi cance. dilution. ANA were detected using anti-mouse IgG conjugated to HRP at a 1:1000 dilution. Top graph: ANA at 4 months post-exposure. Middle Graph: ANA at 7 months post- 3.5. Urine albumin: proteinuria exposure. *Fisher's exact test gave a p < 0.05 for AzA compared to Saline controls. Bottom graph: most of the positive tests were from the female mice at 7 months. Mouse Urine albumin was used as a measure of proteinuria, suggesting numbers were as in Fig. 1. kidney involvement. Fig. 5 indicates that both LAA and AzA were af- fecting kidney function to a moderate degree, possibly by immune not significantly affected by treatment, both fiber types caused an in- complex deposition (Pfau et al., 2008; Zebedeo et al., 2014). Positive crease in the number of B cells in the spleen. The increase in the number tests were indicated by detection of albumin at > 30 mg/dL. There of splenic B cells for AzA was primarily in the female mice in this group, were no significant differences between males and females in any of the where the female mice had significantly more B cells/g mouse than treatment groups (Table 4, Fisher's Exact Test). males (Table 4). Table 4 also shows the percent positive of B cells and T cells separated by sex in each treatment group. With AzA, the percent of 3.6. Spleen lymphocyte subsets by flow cytometry B cells increased, while the percent of T cells decreased in both sexes. Fig. 6, top right, shows the T cell subsets of T helper cells (CD3+, Lymphocyte subsets were evaluated by flow cytometry. Because the CD4+) and T regulatory cells (CD4+, CD25+, FoxP3+). The only spleen weights and cell counts were not statistically different among the significant difference was a reduction in the percentage of T regulatory groups, much of the data is reported as percentage of the parent po- cells with AzA treatment. There were no significant differences between pulation. Fig. 6, top left, shows that while the numbers of T cells were the sexes for these subsets (data not shown).

28 J.C. Pfau et al. Toxicology and Applied Pharmacology 334 (2017) 24–34

difference in the mean percent positive for splenic suppressor B cells in any group. There were also no significant differences between the sexes for these subsets (data not shown).

3.7. Peritoneal lymphocyte subsets

Similarly, lymphocytes in the peritoneal wash fluid were evaluated by flow cytometry. There were no differences among the main populations found in the peritoneal cavity (monocytes and lymphocytes, Fig. 7, top), but among the lymphocytes there were moderate increases in B cells by both LAA and AzA (Fig. 7, bottom). However, consistent with previous studies (Pfau et al., 2013), the percent of lymphocytes that were B1a B cells was reduced with either LAA or AzA (Fig. 7, bottom). Interestingly, male mice contributed more to this effect than females (Table 2). Table 2 shows all of the subpopulation data for each treatment group for both sexes of mice. There was a decrease in the percentage of cells that would be considered suppressor B cells, particularly in the female mice. Overall, this meant that the ratio of B1a to suppressor B cells increased for females with fiber treatment, but decreased for males.

3.8. Serum cytokines by cytokine bead array (CBA)

Table 3 shows the CBA data for cytokines in the serum of the mice. The key ﬁnding was that both LAA and AzA triggered all three of the

cytokines described as part of a TH17 response: TNFα, IL-6, and IL-17. These TH17 cytokines are shown in Fig. 8. Neither LAA or AzA triggered IL-2 or high levels of interferon gamma, which are essential for a TH1 response. Further, neither ﬁber stimulated IL-4, which would have in-

dicated a TH2 response. There were a few diﬀerences between the sexes for LAA only, including IL-10, IFN gamma, and TNFα, where females tended to produce less of these cytokines (Table 3).

3.9. Lung and pleural collagen by hydroxyproline (QuickZyme) assay

Non-cancer pulmonary disease with LAA exposure includes a lim- Fig. 4. MCAA measured by cell-based ELISA. Top graph: Percent positive for MCAA. ited number of cases of interstitial fibrosis, with the predominant dis- * = p < 0.05 by Fisher's Exact Test. Bottom Graph: MCAA scores = number of standard ease being pleural fibrosis (Peipins et al., 2003; Szeinuk et al., 2016; deviations above the mean for the saline controls. Error bars = SEM. * = p < 0.05 Black et al., 2014). In the mice, interstitial fibrosis was measured as compared to Saline controls and LAA. Mouse numbers were as in Fig. 1. total collagen in the right lung of each mouse. Fig. 9, top, demonstrates that excess lung collagen occurred with both LAA and AzA to a similar degree. Female mice had more lung collagen than males in controls, but this did not increase significantly with treatment (Table 5). Significant increases in lung collagen with fiber treatment occurred in the males. Pleural fibrosis was assessed as well, with the diaphragms assayed as parietal pleura. Fig. 9, bottom, shows the diaphragm collagen results, with excess collagen consistently occurring with both LAA and AzA. These increases occurred in both male and female mice, with no difference between males and females (Table 5).

4. Discussion

The revelation that the vermiculite mined for decades outside of Libby, MT, was contaminated with amphibole asbestos has led to multiple studies of the health outcomes of exposure to Libby Amphibole Asbestos (LAA). There remains no doubt that LAA can trigger me- Fig. 5. Urine albumin was measured using Albustix with freshly collected mouse urine. sothelioma, pulmonary carcinoma, interstitial fibrosis (asbestosis), and The percent positive in each treatment group is graphed, where the positive/negative cut- pleural fibrosis (Peipins et al., 2003; Rohs et al., 2008; Sullivan, 2007; off was 30 mg/dL; readings of “trace” albumin were considered negative. * = p < 0.005 Whitehouse et al., 2008; Winters et al., 2012; U.S. Environmental by Fisher's Exact Test. Protection Agency, R, 2015; Black et al., 2014; Antao et al., 2012; Mouse numbers were as in Fig. 1. Bandli & Gunter, 2006; Benson et al., 2015; Dunning et al., 2012; Larson et al., 2010b; Larson et al., 2012a; Larson et al., 2012b; McDonald et al., Fig. 6, bottom, shows the B cell subsets of B1a B cells (IgM 2004; Whitehouse, 2004), consistent with our general understanding of +/CD5+, CD23-) and suppressor B cells (IgM+, CD5+, CD11b+, asbestos toxicity. However, its predominant health outcomes appear to CD1+). Both LAA and AzA led to a significantly greater percentage of be a progressive lamellar pleural thickening (LPT) (Szeinuk et al., 2016; B1a B cells among the lymphocyte population, with no significant Black et al., 2014), and systemic (rheumatic) autoimmune diseases

29 J.C. Pfau et al. Toxicology and Applied Pharmacology 334 (2017) 24–34

Fig. 6. Top left: T (CD3+) and B (CD19+) cells were identiﬁed in the lymphocyte gate by ﬂow cytometry, using isotype controls to determine background values for positive staining. Percent positive was multiplied by the total number of lymphocytes. Top right: T cell subsets. CD3+ is parent for T helper; T helper is parent for Treg. Bottom: B cell subsets. Lymphocytes are parent for B1a; B1a is parent for B suppressors. Error bars = SEM. * = p < 0.01, ** = p < 0.001 compared to Saline control, by 2-tailed t-test and by one-way ANOVA.

Fig. 7. Peritoneal Cell populations by flow cytometry. Top: Lymphocytes and monocytes were identified by Forward vs Side Scatter. Bottom: Lymphocyte subsets were identified by antibodies described in the Materials & Methods. Parent for T, B and B1a cells is Lymphocytes. Parent for B Supp are B1a B cells. Error bars = SEM. * = p < 0.05 compared to Saline controls.

(SAID) (Pfau et al., 2005; Noonan et al., 2006). The latter includes a disease manifestations and severity. rheumatic pain syndrome characterized by positive antinuclear auto- We undertook this study to test the hypothesis that what has been antibody (ANA) tests, joint pain/swelling, and severe fatigue (Un- learned from LAA can be applied to novel “naturally occurring as- published data, Libby Epidemiology Research Program). Mouse studies bestos” (NOA). Such NOA is being discovered in many areas of the U.S. have supported these findings, such that the mouse model has proven to and around the world (Abakay et al., 2016; Bayram & Bakan, 2014; be extremely helpful in understanding some of these mechanisms. A key Buck et al., 2013; Metcalf & Buck, 2015), making it an emerging public finding that was consistent in both mice and humans was that, while health hazard. Due to the severity and latency of asbestos-related dis- amphibole asbestos drives autoantibody production, chrysotile does not eases, identification of exposures and early health risk assessment is appear to do so, suggesting that immune effects could be fiber-specific essential in order to control/reduce exposure as much as possible. (Ferro et al., 2013; Pfau et al., 2015). This may then affect down-stream This study is novel for two reasons. First, it includes a newly

30 J.C. Pfau et al. Toxicology and Applied Pharmacology 334 (2017) 24–34

Table 2 Peritoneal cell subpopulations by ﬂow cytometrya.

Saline LAA AzAA P valueb (by treatments)

Monocytes Male 17.7 (4.5) 23.2 (9.8)c 18.2 (7.2)c 0.39 Female 12.2 (4.1) 11.7 (3.5) 11.3 (4.9) 0.93 Lymphocytes Male 63.1 (5.6) 55.4 (12.2) 60.5 (10.2) 0.37 Female 63.5 (10.6) 67.9 (11.7) 64.4 (6.7) 0.72 T cell Male 15.3 (3.4) 11.8 (4.4) 21.4 (6.4)c 0.01 Female 13.2 (3.1) 10.7 (1.6) 10.6 (2.5) 0.17 B cells Male 49.2 (9.4) 63.2 (2.5) 56.5 (7.6) 0.025 Female 59.3 (6.6) 66.9 (10.7) 61.6 (8.3) 0.37 B1a B cells Male 26.1 (7.1) 12.5 (3.5) 18.1 (7.4) 0.009 Female 19.9 (5.0) 13.6 (5.5) 19.3 (7.4) 0.23 B supp

Male 4.9 (1.7) 6.0 (3.1) 4.1 (1.6) 0.34 Fig. 8. Serum Cytokines by CBA. Cytokines of the TH17 proﬁle are shown. Error Female 6.9 (3.2) 3.3 (1.5)c 2.0 (1.2)c 0.002 bars = SEM. a = p < 0.01 compared to Saline control, b = p < 0.05 compared to LAA. a Data shown as mean percent of parent population (std deviation). b One Way ANOVA with post-hoc tests; bold = p < 0.05 compared to saline control. c p < 0.05 compared to other sex. Table 4 Male/female data comparison – spleen data.

Table 3 Saline LAA AzA Serum cytokines by cytokine bead arraya.

b Spleen Weight, mg (SEM) Saline LAA AzA p value treatment Male 94.4 (6.4) 92.3 (2.5) 95.1 (4.8) Female 98.8 (6.1) 116.0 128.7 (12.6) IL-4 11.6 (0.37) 11.8 (0.32) 13.3 (2.2) > 0.05 (13.4) Male 11.5 (0.35) 11.9 (0.47) 12.6 (0.47) > 0.05 Splenocytes/g mouse, millions (SEM) Female 11.9 (0.15) 11.7 (0.15) 13.8 (3.0) > 0.05 Male 1.55 (0.3) 2.36 (0.3) 1.08 (0.4) IL-10 69.0 (8.6) 117.6 (54.5) 169.8 (30.6) < 0.01 Female 1.48 (0.1) 1.66 (0.2) 3.26 (0.5)a,b Male 65.2 (8.0) 163.6 (11.5) 185.3 (25.6) < 0.01 Spleen B cells/g mouse, millions (SEM) Female 76.0 (4.0) 80.8 (11.5)c 156.1 (28.0) < 0.01 Male 0.24 0.56 (0.2) 0.38 (0.1) IL-2 7.0 (0.29) 6.9 (0.32) 6.7 (0.55) > 0.05 (0.04) Male 7.0 (0.27) 6.6 (0.53) 6.8 (0.39) > 0.05 Female 0.19 0.33 (0.1) 0.93 (0.2)a,b Female 7.1 (0.26) 7.3 (0.28) 6.6 (0.40) > 0.05 (0.04) IFN gamma 4.0 (0.77) 9.1 (5.2) 10.9 (6.2) < 0.05 Spleen T cells/g mouse, millions (SEM) Male 3.8 (0.64) 14.1 (2.4) 9.1 (1.9) < 0.05 Male 0.39 0.47 (0.21) 0.35 (0.14) Female 4.3 (0.96) 5.0 (0.98)c 12.5 (8.3) > 0.05 (0.11) TNF alpha 25.6 (2.2) 41.6 (18.2) 64.2 (11.1) < 0.05 Female 0.39 0.30 (0.05) 0.41 (0.09) Male 25.1 (1.8) 56.8 (18.0) 62.5 (8.8) < 0.05 (0.04) Female 26.6 (2.7) 29.4 (2.4)c 65.7 (13.2) < 0.05 Spleen B cells, mean percent of IL-6 8.3 (1.1) 10.5 (2.6) 20.2 (6.7) < 0.05 lymphocytes (SEM) Male 7.9 (0.96) 12.2 (3.1) 18.0 (1.0) < 0.05 Male 43.4 (4.4) 58.0 (8.8) 73.4 (2.5)b Female 8.9 (1.1) 9.2 (1.1) 22.1 (9.0) < 0.05 Female 31.5 (4.6) 52.7 (6.0) 70.5 (4.6)b IL-17 1.4 (0.95) 12.6 (2.5) 9.7 (3.1) < 0.01 Spleen T cells, mean percent of Male 1.1 (0.88) 10.8 (2.6) 8.6 (1.4) < 0.01 lymphocytes (SEM) Female 1.9 (0.94) 14.0 (1.4) 10.7 (3.9) < 0.01 Male 56.7 (4.3) 42.0 (7.2) 26.6 (2.5)b Female 68.5 (4.6) 47.3 (6.0) 29.5 (4.6)b a Data shown as mean concentration in pg/ml (std deviation). b One Way ANOVA with post-hoc tests; bold = p < 0.05 compared to saline control. a 2-Tailed, unpaired t-test, Bold: p < 0.05 compared to other sex. c p < 0.05 compared to other sex. b One Way ANOVA with post-hoc tests; Bold = p < 0.05 compared to saline control. fi described amphibole ber mixture that has never been evaluated for et al., 2016). Taking into account rat respiratory rate and volume per ff health e ects. Exposures to this material are occurring due to (a) day, and then calculating a total exposure per gram of rat, the total dose fi widespread geographic occurrence of the brous amphiboles across for the lowest exposure level was 3.74 μg/g of rat. This is comparable to southern Nevada and Arizona (Buck et al., 2013; Metcalf & Buck, 2015), our previous exposure dose of 60 μg/mouse (Ferro et al., 2013), cal- (b) highly popular recreation sites in contaminated areas, creating large culated as 2.4 μg/g of mouse (considering a 25 g mouse). Another study amounts of dust, (c) urban development expanding into and currently used a mass dose per rat of 0.5 or 1.5 mg/rat, for which the lowest dose developed within the contaminated areas, and (d) the arid climate of calculates to 1.2 μg/g of rat (Cyphert et al., 2012). Our current dose of this region that promotes naturally-generated asbestos-laden dust. 3 μg/mouse is 1/10th of even that low dose, calculating at 0.12 μg/g of Second, whereas we have used this mouse model to extensively study mouse. It is therefore, to our knowledge, the lowest dose ever used in ff – μ e ects of LAA, we consistently used an exposure dose of 60 120 g/ rodent models of asbestos exposure. The purpose for this, as mentioned mouse (Ferro et al., 2013; Pfau et al., 2008), closer to what could above in the Methods, is that environmental exposures tend to be correspond to a high cumulative occupational exposure. The current sporadic, due to dust storms or recreational activity in areas with μ study used only 3 g/mouse, closer to what might occur as a periodic naturally occurring asbestos, rather than regular, all day, work ex- fi environmental exposure. Other ber comparison studies provide some posures. The cumulative exposure would thus be expected to be very perspective on this dose: Gavett, et al., exposed rats to LAA in air at 0.5, fi ff 3 low. The fact that signi cant e ects were seen in the mice at the dose 3.5, and 25 mg/m for 6 h/day, 5 days/week for 13 weeks (Gavett used in this study raises serious concern about such environmental

31 J.C. Pfau et al. Toxicology and Applied Pharmacology 334 (2017) 24–34

Table 5 morphology because different techniques were used to characterize the Other male/female data comparisons. fibers (Table 1). The TEM analyses used for LAA allow for smaller particles to be imaged and analyzed as compared to the SEM analyses Saline LAA AzA used for AzA. These differences likely explain much of the differences Urine albumin, % positive (# shown in Table 1 between AzA and LAA. However, overall, LAA has positive) lower aspect ratios as compared to AzA both in the mean (8.4 ± 0.7 Male 22.2 (2 of 9) 100 (4 of 4) 100 (7 of 7) for LAA vs. 16.7 ± 0.9 for AzA whole sample including accessory Female 0.0 (0 of 5) 75 (3 of 4) 63 (5 of 8) Lung collagen, μg/mg lung (SEM) minerals) as well as the percentage of particles with aspect ratios 3.0 or Male 1.2 (0.2) 2.45 (0.05)b 2.4 (0.1)b greater (49% for LAA; 86% for AzA). Female 2.1 (0.15)a 2.3 (0.14) 2.4 (0.07) Diaphragm collagen, μg/mg (SEM) Male 31.3 (7.1) 67.7 (8.0)b 98.5 (17.9)b 4.1. Immune/inflammatory effects Female 27.5 (13.1) 62.8 (20.7)b 65.6 (13.1)b MCAA, % positive (# positive) Male 11.1 (1 of 9) 25 (1 of 4) 50 (4 of 8) In this study, AzA is at least as immunostimulatory as LAA. A higher Female 20 (1 of 5) 40 (2 of 5) 66 (6 of 9) dose LAA (60–120 μg/mouse) leads to 80% of mice having positive MCAA score, mean (SEM) ANA tests (Ferro et al., 2013; Pfau et al., 2008; Zebedeo et al., 2014). Male 1.24 (0.79) 1.68 (1.32) 2.98 (0.81) Female 1.76 (0.67) 1.74 (1.33) 4.88 (0.97) However, at the current low dose, only AzA induced ANA above control frequencies. In the previous high dose studies, not only was the dose a 2-Tailed, unpaired t-test, Bold: p < 0.05 compared to other sex. higher than what was used here, but only female mice were used (Ferro b One Way ANOVA with post-hoc tests; Bold = p < 0.05 compared to saline control. et al., 2013; Pfau et al., 2008). The current study shows that female mice are more likely to produce positive ANA tests; therefore, the in- clusion of male mice in the current study affected the overall results. Another difference was that previous studies used intratracheal instillation, which likely affects both the size and quantity of fibers reaching the lower respiratory tract, compared to oropharyngeal aspiration used here, by bypassing the pharynx and upper trachea. A recent comparison of intratracheal, intranasal, and oropharyngeal aspiration of crystalline silica particles demonstrated comparable effects, but intratracheal instillation led to more dispersion of the particles throughout the lung and more potent inflammatory effects (Lacher et al., 2010). We used oropharyngeal aspiration in the current study to avoid surgical and pharmaceutical stressors, and to more closely model human exposures for these fibers. This method has been shown to be somewhat more effective at recapitulating outcomes seen in humans compared to intratracheal (Lacher et al., 2010; Egger et al., 2013; Lakatos et al., 2006). Despite these differences, immune dysfunction was apparent in both male and female mice at the current low dose of both LAA and AzA, including a shift in lymphocyte subsets in both spleen and peritoneal cavity. In the spleen, the percentage of regulatory T cells was slightly reduced despite no overall difference in spleen weight, lymphocytes, or overall T cell numbers. This is consistent with our previous study showing reduced numbers of regulatory T cells in lymph nodes of amphibole-exposed mice (Pfau et al., 2008). Total spleen B cell numbers were increased with both LAA and AzA, and among those B cells, the proportion of B1a B cells was increased, again consistent with previous studies with LAA (Ferro et al., 2013; Pfau et al., 2013). B1a B cells have been implicated in autoantibody production and high numbers are seen in lupus-prone mouse strains (Duan & Morel, 2006; Mohan et al., 1998; Potula et al., 2012). Unlike our previous studies, there was no change in the frequency of splenic suppressor B cells with any treatment. However, AzA did cause a reduction in the frequency of both peritoneal B1a B cells and suppressor B cells, again consistent with previous studies for LAA (Ferro et al., 2013; Pfau et al., 2013). This Fig. 9. Lung and Pleural Collagen by hydroxyproline assay (QuickZyme). Top: Total suggests that activation and trafficking of B1a B cells to the spleen were collagen in the right lung of each mouse was assayed and data are shown as μg of collagen affected. This is a key observation as the spleen is the site where these per mg of wet lung tissue. ** = p < 0.01 compared to saline control. Bottom: Collagen cells become active in producing autoantibodies (Pfau et al., 2013; in diaphragms (Parietal pleura). * = p < 0.01 compared to Saline. Mohan et al., 1998).

Both LAA and AzA increased serum cytokines of the TH17 response exposures. profile, but not TH1orTH2 responses. This is consistent with our pre- Using C57BL/6 mice, we compared the immune and pulmonary vious high doses studies with LAA as well (Ferro et al., 2013; Zebedeo ff fi e ects of this low level exposure to LAA and a mixture of brous am- et al., 2014). TH17 responses have been implicated in autoimmune and phibole asbestos from Arizona (AzA). The AzA is composed of winchite inflammatory diseases (Afzali et al., 2007; Furuzawa-Carballeda et al., (76%) and actinolite (24%), which has very similar mineralogy to LAA, 2007; Kim et al., 2016). Both males and females contributed to the composed of winchite (83%), richterite (11%), and tremolite (9%) increases in TH17 cytokines for both fibers. (Meeker et al., 2003)(Table 1). It is more difficult to compare particle

32 J.C. Pfau et al. Toxicology and Applied Pharmacology 334 (2017) 24–34

4.2. Pulmonary effects condition and distance from naturally occurring asbestos. Environ. Health Prev. Med. 21 (2), 82–90. Afzali, B., Lombardi, G., Lechler, R.I., Lord, G.M., 2007. The role of T helper 17 (Th17) Both LAA and AzA induced interstitial and pleural fibrosis, which and regulatory T cells (Treg) in human organ transplantation and autoimmune dis- are well-known outcomes of asbestos exposure. In humans, the unique ease. Clin. Exp. Immunol. 148 (1), 32–46. Antao, V.C., Larson, T.C., Horton, D.K., 2012. Libby vermiculite exposure and risk of LPT of LAA is thought to occur primarily on the parietal pleura, but to developing asbestos-related lung and pleural diseases. Curr. Opin. Pulm. Med. 18 (2), also involve the visceral pleura (Szeinuk et al., 2016). There were no 161–167. significant differences between LAA and AzA in terms of the quantity of Bandli, B.R., Gunter, M.E., 2006. A review of scientific literature examining the mining ff collagen detected in lungs or diaphragms at this low dose. Untreated history, geology, mineralogy, and amphibole asbestos health e ects of the Rainy Creek igneous complex, Libby, Montana, USA. Inhal. Toxicol. 18 (12), 949–962. female mice had more lung collagen than males, and responded mini- Baumann, F., Ambrosi, J.P., Carbone, M., 2013. Asbestos is not just asbestos: an un- mally to treatment. However, male mice had the significant increases in recognised health hazard. Lancet Oncol. 14 (7), 576–578. lung collagen with both fibers. Baumann, F., Buck, B.J., Metcalf, R.V., McLaurin, B.T., Merkler, D.J., Carbone, M., 2015. The presence of asbestos in the natural environment is likely related to mesothelioma In humans and mice, there is an association between the presence of in young individuals and women from southern Nevada. J. Thorac. Oncol. 10 (5), mesothelial cell autoantibodies (MCAA) and pleural fibrosis (Marchand 731–737. et al., 2012; Gilmer et al., 2015). This is a critical finding due to the Bayram, M., Bakan, N.D., 2014. Environmental exposure to asbestos: from geology to mesothelioma. Curr. Opin. Pulm. Med. 20 (3), 301–307. severe and progressive nature of the amphibole-induced LPT, since Benson, R., Berry, D., Lockey, J., Brattin, W., Hilbert, T., LeMasters, G., 2015. Exposure- MCAA may prove to be a therapeutic target for slowing the progression response modeling of non-cancer effects in humans exposed to Libby Amphibole or severity of disease (Gilmer et al., 2016; Hanson et al., 2016). Asbestos; update. Regul. Toxicol. Pharmacol. 73 (3), 780–789. fi Black, B., Szeinuk, J., Whitehouse, A.C., Levin, S.M., Henschke, C.I., Yankelevitz, D.F., Therefore, the induction of both MCAA and pleural brosis in this Flores, R.M., 2014. Rapid progression of pleural disease due to exposure to Libby study, along with the ability of MCAA to induce serous fibrosis in mice amphibole: “not your grandfather's asbestos related disease”. Am. J. Ind. Med. 57 in the absence of asbestos (Gilmer et al., 2015), provides evidence and (11), 1197–1206. fi Blake, D.J., Wetzel, S.A., Pfau, J.C., 2008. Autoantibodies from mice exposed to Libby justi cation for further study into the pathogenic role of MCAA. amphibole asbestos bind SSA/Ro52-enriched apoptotic blebs of murine macrophages. Toxicology 246 (2–3), 172–179. 4.3. Limitations and conclusions Buck, B.J., Goossens, D., Metcalf, R.V., McLaurin, B., Ren, M., Freudenberger, F., 2013. Naturally occurring asbestos: potential for human exposure, southern Nevada, USA. Soil Sci. Soc. Am. J. 77, 2192–2204. The limitations of this study include the use of non-elutriated fibers Carbone, M., Kanodia, S., Chao, A., Miller, A., Wali, A., Weissman, D., Adjei, A., and differences in fiber characterization methods. Elutriation is a Baumann, F., Boffetta, P., Buck, B., de Perrot, M., Dogan, A.U., Gavett, S., Gualtieri, ff method of collecting fibers, from a raw sample, that fall into a respir- A., Hassan, R., Hesdor er, M., Hirsch, F.R., Larson, D., Mao, W., Masten, S., Pass, H.I., fi Peto, J., Pira, E., Steele, I., Tsao, A., Woodard, G.A., Yang, H., Malik, S., 2016. able range, meaning that the bers would be expected to reach the Consensus report of the 2015 Weinman international conference on mesothelioma. J. small bronchioles and alveoli of the lung (Webber et al., 2008). Non- Thorac. Oncol. 11 (8), 1246–1262. fi Cooper, W.C., Murchio, J., Popendorf, W., Wenk, H.R., 1979. Chrysotile asbestos in a elutriated bers were used in order to set the baseline in mouse studies – fi California recreational area. Science 206 (4419), 685 688. where the full range of ber sizes in collected samples was included Cyphert, J.M., Nyska, A., Mahoney, R.K., Schladweiler, M.C., Kodavanti, U.P., Gavett, (columns 1 and 3 in Table 1). Oropharyngeal exposures tend to leave S.H., 2012. Sumas Mountain chrysotile induces greater lung fibrosis in Fischer344 larger particles and fiber bundles in the pharynx and upper airways, rats than Libby amphibole, El Dorado tremolite, and Ontario ferroactinolite. Toxicol. – fi Sci. 130 (2), 405 415. allowing the body to naturally sort a mixture of bers into lung com- Duan, B., Morel, L., 2006. Role of B-1a cells in autoimmunity. Autoimmun. Rev. 5 (6), partments. Additionally, the different techniques available to char- 403–408. acterize fibers have their own limitations. Transmission electron mi- Duncan, K.E., Cook, P.M., Gavett, S.H., Dailey, L.A., Mahoney, R.K., Ghio, A.J., Roggli, V.L., Devlin, R.B., 2014. In vitro determinants of asbestos fiber toxicity: effect on the croscopy (TEM/SAED) allows for analysis of smaller particles compared relative toxicity of Libby amphibole in primary human airway epithelial cells. Part to SEM/EDS. Future studies using field-emission SEM may be able to Fibre Toxicol. 11, 2. better capture these smaller particles, and surface area analyses can also Dunning, K.K., Adjei, S., Levin, L., Rohs, A.M., Hilbert, T., Borton, E., Kapil, V., Rice, C., be used to better understand what characteristics might be driving the Lemasters, G.K., Lockey, J.E., 2012. Mesothelioma associated with commercial use of vermiculite containing Libby amphibole. J. Occup. Environ. Med. 54 (11), immune dysfunction. 1359–1363. In conclusion, these data support our hypothesis that other en- Egger, C., Cannet, C., Gerard, C., Jarman, E., Jarai, G., Feige, A., Suply, T., Micard, A., vironmental fibrous amphiboles can have effects very much like LAA, Dunbar, A., Tigani, B., Beckmann, N., 2013. Administration of bleomycin via the oropharyngeal aspiration route leads to sustained lung fibrosis in mice and rats as even at very low doses in mice. This should raise the level of concern quantified by UTE-MRI and histology. PLoS One 8 (5), e63432. regarding exposures in southern Nevada, Arizona, and other places Environmental Protection Agency, US, 2008. Clear Creek Management Area Asbestos where natural processes and anthropogenic activities are generating Exposure And Human Health Risk Assessment, R. 9, Editor. U.S. EPA, San Franciso, CA. dust. Ferro, A., Zebedeo, C.N., Davis, C., Ng, K.W., Pfau, J.C., 2013. Amphibole, but not chrysotile, asbestos induces anti-nuclear autoantibodies and IL-17 in C57BL/6 mice. Funding J. Immunotoxicol. http://dx.doi.org/10.3109/1547691X.2013.847510. Furuzawa-Carballeda, J., Vargas-Rojas, M.I., Cabral, A.R., 2007. Autoimmune inflammation from the Th17 perspective. Autoimmun. Rev. 6 (3), 169–175. This work was funded by a grant from the Montana Agricultural Gavett, S.H., Parkinson, C.U., Willson, G.A., Wood, C.E., Jarabek, A.M., Roberts, K.C., Experimental Station (MAES), Montana State University, Bozeman, MT. Kodavanti, U.P., Dodd, D.E., 2016. Persistent effects of Libby amphibole and amosite asbestos following subchronic inhalation in rats. Part Fibre Toxicol. 13, 17. Gilmer, J., Serve, K., Davis, C., Anthony, M., Hanson, R., Harding, T., Pfau, J.C., 2015. Declarations of interest Libby amphibole induced mesothelial cell autoantibodies promote collagen deposition in mice. Am. J. Phys. Lung Cell. Mol. Phys. 2016 p. ajplung 00462. The authors have no conflicts of interest to disclose. Gilmer, J., Harding, T., Woods, L., Black, B., Flores, R., Pfau, J.C., 2016. Mesothelial cell autoantibodies upregulate transcription factors associated with fibrosis. Inhal. Toxicol (in press). Acknowledgements Hanson, R., Evilia, C., Gilmer, J., Woods, L., Black, B., Flores, R., Pfau, J.C., 2016. Libby amphibole-induced mesothelial cell autoantibodies bind to surface plasminogen and – fi alter collagen matrix remodeling. Phys. Rep. 4 (7), 1 12. We thank Isabella Aquino for assistance in collecting ber char- Kim, B.S., Park, Y.J., Chung, Y., 2016. Targeting IL-17 in autoimmunity and inflamma- acterization data. tion. Arch. Pharm. Res. 39 (11), 1537–1547. Lacher, S.E., Johnson, C., Jessop, F., Holian, A., Migliaccio, C.T., 2010. Murine pulmonary inflammation model: a comparative study of anesthesia and instillation methods. References Inhal. Toxicol. 22 (1), 77–83. Lakatos, H.F., Burgess, H.A., Thatcher, T.H., Redonnet, M.R., Hernady, E., Williams, J.P., Abakay, A., Tanrikulu, A.C., Ayhan, M., Imamoglu, M.S., Taylan, M., Kaplan, M.A., Sime, P.J., 2006. Oropharyngeal aspiration of a silica suspension produces a superior Abakay, O., 2016. High-risk mesothelioma relation to meteorological and geological model of silicosis in the mouse when compared to intratracheal instillation. Exp. Lung

33 J.C. Pfau et al. Toxicology and Applied Pharmacology 334 (2017) 24–34

Res. 32 (5), 181–199. Asbestos-induced autoimmunity in C57BL/6 mice. J. Immunotoxicol. 5 (2), 129–137. Larson, T.C., Meyer, C.A., Kapil, V., Gurney, J.W., Tarver, R.D., Black, C.B., Lockey, J.E., Pfau, J.C., Hurley, K., Peterson, C., Coker, L., Fowers, C., Marcum, R., 2013. Activation 2010a. Workers with Libby amphibole exposure: retrospective identification and and trafficking of peritoneal B1a B-cells in response to amphibole asbestos. J. progression of radiographic changes. Radiology 255 (3), 924–933. Immunotoxicol (in press). Larson, T.C., Antao, V.C., Bove, F.J., 2010b. Vermiculite worker mortality: estimated Pfau, J.C., Serve, K.M., Woods, L., Noonan, C.W., 2015. Asbestos Exposure and effects of occupational exposure to Libby amphibole. J. Occup. Environ. Med. 52 (5), Autoimmunity, in Biological Effects of Fibrous and Particulate Substances. Springer 555–560. Publications. Larson, T.C., Antao, V.C., Bove, F.J., Cusack, C., 2012a. Association between cumulative Potula, H.H., Xu, Z., Zeumer, L., Sang, A., Croker, B.P., Morel, L., 2012. Cyclin-dependent fiber exposure and respiratory outcomes among Libby vermiculite workers. J. Occup. kinase inhibitor Cdkn2c deficiency promotes B1a cell expansion and autoimmunity in Environ. Med. 54 (1), 56–63. a mouse model of lupus. J. Immunol. 189 (6), 2931–2940. Larson, T.C., Lewin, M., Gottschall, E.B., Antao, V.C., Kapil, V., Rose, C.S., 2012b. Rohs, A.M., Lockey, J.E., Dunning, K.K., Shukla, R., Fan, H., Hilbert, T., Borton, E., Wiot, Associations between radiographic findings and spirometry in a community exposed J., Meyer, C., Shipley, R.T., Lemasters, G.K., Kapil, V., 2008. Low-level fiber-induced to Libby amphibole. Occup. Environ. Med. 69 (5), 361–366. radiographic changes caused by Libby vermiculite: a 25-year follow-up study. Am. J. Lockey, J.E., Dunning, K., Hilbert, T.J., Borton, E., Levin, L., Rice, C.H., McKay, R.T., Respir. Crit. Care Med. 177 (6), 630–637. Shipley, R., Meyer, C.A., Perme, C., LeMasters, G.K., 2015. HRCT/CT and associated Serve, K.M., Black, B., Szeinuk, J., Pfau, J.C., 2013. Asbestos-associated mesothelial cell spirometric effects of low Libby amphibole asbestos exposure. J. Occup. Environ. autoantibodies promote collagen deposition in vitro. Inhal. Toxicol. 25 (14), Med. 57 (1), 6–13. 774–784. Lowers, H.A., Wilson, S.A., Hoefen, T.M., Benzel, W.M., Meeker, G.P., 2012. Preparation Sullivan, P.A., 2007. Vermiculite, respiratory disease, and asbestos exposure in Libby, and Characterization of “Libby Amphibole” Toxicological Testing Material, D.o.t. Montana: update of a cohort mortality study. Environ. Health Perspect. 115 (4), Interior, Editor. United States Geological Survey, Reston, VA, pp. 1–20. 579–585. Marchand, L.S., St-Hilaire, S., Putnam, E.A., Serve, K.M., Pfau, J.C., 2012. Mesothelial cell Szeinuk, J., Noonan, C.W., Henschke, C.I., Pfau, J., Black, B., Miller, A., Yankelevitz, D.F., and anti-nuclear autoantibodies associated with pleural abnormalities in an asbestos Liang, M., Liu, Y., Yip, R., Linker, L., McNew, T., Flores, R.M., 2016. Pulmonary exposed population of Libby MT. Toxicol. Lett. 208 (2), 168–173. abnormalities as a result of exposure to Libby amphibole during childhood and McDonald, J.C., Harris, J., Armstrong, B., 2004. Mortality in a cohort of vermiculite adolescence-the pre-adult latency study (PALS). Am. J. Ind. Med. miners exposed to fibrous amphibole in Libby, Montana. Occup. Environ. Med. 61 Tetra Tech, I, 2014. Phase I Site Characterization Report for Boulder City Bypass (4), 363–366. Naturally Occurring Asbestos (NOA) Project: Railroad Pass to Silverline Road. Meeker, G.P., Bern, A.M., Brownfield, I.K., Lowers, H.A., Sutley, S.J., Hoefen, T.M., Nevada Department of Transportation, Helena MT, pp. 1–118. Vance, J.S., 2003. The composition and morphology of amphiboles from the Rainy U.S. Environmental Protection Agency, R, 2011. Libby Asbestos Superfund Site Operable Creek complex, near Libby, Montana. Am. Mineraologist 88, 1955–1969. Unit 3 Initial Screening Level Human Health Risk Assessment for Exposure to Metcalf, R.V., Buck, B.J., 2015. Genesis and health risk implications of an unusual oc- Asbestos. US EPA. currence of fibrous NaFe3+-amphibole. Geology 43 (1), 63–66. U.S. Environmental Protection Agency, R, 2015. Site-wide Human Health Risk Mohan, C., Morel, L., Yang, P., Wakeland, E.K., 1998. Accumulation of splenic B1a cells Assessment - Executive Summary - Libby Asbestos Superfund Site, Libby, Montana, R. with potent antigen-presenting capability in NZM2410 lupus-prone mice. Arthritis 8, Editor. CDM Smith. Rheum. 41 (9), 1652–1662. Van Gosen, B.S., 2007. The geology of asbestos in the United States and its practical Noonan, C.W., 2006. Exposure matrix development for the Libby cohort. Inhal. Toxicol. applications. Environ. Eng. Geosci. 13 (1), 55–68. 18 (12), 963–967. Webber, J.S., Blake, D.J., Ward, T.J., Pfau, J.C., 2008. Separation and characterization of Noonan, C.W., Pfau, J.C., Larson, T.C., Spence, M.R., 2006. Nested case-control study of respirable amphibole fibers from Libby, Montana. Inhal. Toxicol. 20 (8), 733–740. autoimmune disease in an asbestos-exposed population. Environ. Health Perspect. Whitehouse, A.C., 2004. Asbestos-related pleural disease due to tremolite associated with 114 (8), 1243–1247. progressive loss of lung function: serial observations in 123 miners, family members, Noonan, C.W., Conway, K., Landguth, E.L., McNew, T., Linker, L., Pfau, J., Black, B., and residents of Libby, Montana. Am. J. Ind. Med. 46 (3), 219–225. Szeinuk, J., Flores, R., 2015. Multiple pathway asbestos exposure assessment for a Whitehouse, A.C., Black, C.B., Heppe, M.S., Ruckdeschel, J., Levin, S.M., 2008. superfund community. J. Expo. Sci. Environ. Epidemiol. 25 (1), 18–25. Environmental exposure to Libby asbestos and mesotheliomas. Am. J. Ind. Med. 51 Paoletti, L., Batisti, D., Bruno, C., Di Paola, M., Gianfagna, A., Mastrantonio, M., Nesti, M., (11), 877–880. Comba, P., 2000. Unusually high incidence of malignant pleural mesothelioma in a Winters, C.A., Hill, W.G., Rowse, K., Black, B., Kuntz, S.W., Weinert, C., 2012. Descriptive town of eastern Sicily: an epidemiological and environmental study. Arch. Environ. analysis of the respiratory health status of persons exposed to Libby amphibole as- Health 55 (6), 392–398. bestos. BMJ Open 2 (6). Peipins, L.A., Lewin, M., Campolucci, S., Lybarger, J.A., Miller, A., Middleton, D., Weis, Wylie, A.G., Candela, P.A., 2015. Methodologies for determining the sources, character- C., Spence, M., Black, B., Kapil, V., 2003. Radiographic abnormalities and exposure to istics, distribution, and abundance of asbestiform and nonasbestiform amphibole and asbestos-contaminated vermiculite in the community of Libby, Montana, USA. serpentine in ambient air and water. J. Toxicol. Environ. Health B Crit. Rev. 18 (1), Environ. Health Perspect. 111 (14), 1753–1759. 1–42. Pfau, J.C., Sentissi, J.J., Weller, G., Putnam, E.A., 2005. Assessment of autoimmune re- Zebedeo, C.N., Davis, C., Pena, C., Ng, K.W., Pfau, J.C., 2014. Erionite induces production sponses associated with asbestos exposure in Libby, Montana, USA. Environ. Health of autoantibodies and IL-17 in C57BL/6 mice. Toxicol. Appl. Pharmacol. 275 (3), Perspect. 113 (1), 25–30. 257–264. Pfau, J.C., Sentissi, J.J., Li, S., Calderon-Garciduenas, L., Brown, J.M., Blake, D.J., 2008.